Cap3 - MKT 2022-11-18 13 - 56 - 53

3 Marketing research
All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.
Marketing research1 is the process of collecting, analysing and interpreting data to inform marketing
decisions. This chapter will introduce a number of techniques of marketing research which can help arts
organizations to gather and use information more efficiently and effectively. This will include sections on:
The benefits and scope of marketing research

Setting research objectives
Research methodologies
Sources of secondary data
Methods for collecting primary data
Conducting a survey
This chapter is more than just a step-by-step guide to the process of research. It encourages the con-
sideration of both the information needs of an organization and the way in which information can be used,
as well as examining the variety of techniques for gathering that information.
The benefits of marketing research

Organizations that have taken marketing research seriously report significant benefits. Many successful
cultural organizations place marketing research at the heart of their planning, informing a wide range of
activities, including programming and marketing decisions, merchandising policy, income generation
strategy and customer service. For example, integrated cross-functional research enables organizations
like the Rotterdam Philharmonic to capture and analyse data on its audiences to get ‘inside the donor’s
head’ and make the most effective approach (Gribling, 2014); while Warwick Arts Centre has built
primary research into its day-to-day activities to understand the ongoing audience experience of its venue
and events. (Anderson, 2013).
Objections to marketing research

Arts organizations may give a number of reasons for not doing marketing research, but the following are
perhaps the most common:
‘ : : : it’s too expensive’

Copyright 2017. Routledge.
A popular misconception is that marketing research is a highly technical discipline which can be suc-
cessfully pursued only by experts (and normally at great expense). Certainly it is a systematic and orderly
process, and it is possible to use sophisticated information technology to interpret large quantities of data.
EBSCO Publishing : eBook Community College Collection (EBSCOhost) - printed on 11/11/2022 12:58 PM via ERASMUS UNIVERSITEIT ROTTERDAM
AN: 1614623 ; Liz Hill, Catherine O'Sullivan, Terry O'Sullivan, Brian Whitehead.; Creative Arts Marketing
Account: s3776028.main.ehost
Marketing research 57
There are some industries and organizations for which vast consumer surveys and detailed statistical
analysis can be extremely cost-effective; but there are many more organizations that reap significant
rewards from a whole host of far less expensive methods for gathering the information that they need to help
them better serve their customers. Even the smallest arts organization can benefit from a more rigorous
analysis of its box office data, informal discussions with its regular attenders and a closer examination of
the impact of press comment and publicity material, all of which can be done at very little cost.
To say that marketing research is too expensive implies that an organization may not have fully
considered the range of marketing data available and the very limited costs involved in the majority of
data collection. Perhaps, also, it has not considered the cost of poor decision making which could lead to
marketing mistakes and consequently lost audiences.
‘ : : : we don’t want to be popular’
It is sometimes feared that marketing research will identify demand for a programme of events which
would be popular but lacking in artistic merit. Moriarty (1997) summed up the wariness arts workers have
towards evaluation as “the anxiety that something very precious may be lost, that the complexity of an
experience which includes relationship, enjoyment, learning, exploration, expression will be destroyed,
diluted or reduced”. The research may be resisted on the grounds that its findings would not be
implemented for cultural or artistic reasons. While accessible, demand-driven programming is supported
in some circles, and seen as a commitment by more progressive directors to reaching as wide an audience
as possible, by others it is condemned as an insult to the serious performers and a betrayal of the classical
culture that the funding bodies intend to be served. Again this is a flimsy reason for ignoring marketing
research, which can be a very useful tool for supporting decisions on programming, facilities and services
within the ruling artistic policy, and without alienating existing interest groups and audiences.
‘ : : : we know what our audiences want’
Commercial arts organizations tend to be keenly aware of their customers’ opinions, which are critical to
their survival, but service providers in the not-for-profit sector can be overconfident in their belief that
they know instinctively what their customers want. These preconceptions can hinder their attempts to
Box 3.1 Quality Metrics

To be eligible for funding, larger cultural organizations in England are to be asked to conduct a
regular audience survey using a standardized set of common questions known as Quality Metrics to
provide a measure of the artistic quality of their programme. Arts Council England believes this
will enable the quality of work being produced by these organizations to be compared across
different art forms and types of organization.
The arts sector reacted with horror at the announcement and took to Twitter to express its
overwhelming disapproval, with comments like “You can’t measure artistic ‘quality’. You can feel
it, and think about it, and criticize it and even assess it. But you can’t measure it”, and “How on earth
can ‘quality metrics’ ever be a meaningful measure of artistic quality? What room does that leave
for innovation?” One blogger even imagined how Michelangelo’s Sistine Chapel might have been
rated under the standardized assessment criteria, concluding that it wouldn’t have ticked all the
boxes!
Sources: ArtsProfessional, 2016; Ward, 2016; Higgins, 2016.
EBSCOhost - printed on 11/11/2022 12:58 PM via ERASMUS UNIVERSITEIT ROTTERDAM. All use subject to https://www.ebsco.com/terms-of-use
achieve their non-profit objectives, especially if they wish to attract a more diverse audience, or reach out
to non-typical attenders and visitors. Marketing research is a key tool to help them understand the
motivations and attitudes of their target markets, and deliver appropriate programmes and experiences.
To reject it is to reject the core of the marketing concept, which places the needs of the customer at the
heart of the organization.
The scope of marketing research

The definition given of marketing research is deliberately broad in scope. It refers to any attempt to gather
information which may be useful in the planning of marketing activity. However, marketing research can
be subdivided into a number of categories, according to its purpose:
Audience profiling
In the arts, this is primarily concerned with understanding the nature, composition and preferences of
current and potential audiences. It is commonly used to help organizations identify audience groupings
(or market segments) with similar characteristics and arts preferences. This enables visitor or audience
profiles to be constructed for different types of exhibitions, performances or events. Profiling an audience
typically involves gathering details of demographic characteristics including age and gender, but also
income, occupation, education, and perhaps means of transport and distance travelled to the venue. This
type of information helps cultural organizations to target their future programmes, events, promotional
literature, fundraising and advertising more precisely. If quantified, it can provide particularly useful
information for negotiating with potential sponsors, who will want to know what sort of people an
organization attracts.
Customer satisfaction research

This type of research aims to measure the extent to which an arts event has met its audiences’ expec-
tations, to enable organizations to understand and quickly respond to their customers’ perceptions of both
the artistic product and the environment in which it is produced. Social media and review websites, such
as TripAdvisor®, offer audiences a public forum to share their views of a cultural organization or event
they have experienced. Clearly this can be a valuable source of information, but may generate unwanted
publicity if a production fails to please its audience. Ticket buyers can be encouraged to give a more
private post-performance critique through a survey delivered to mobile phones immediately after a show.
Motivation research
This attempts to get to the bottom of audiences’ reasons for attending a particular event or venue. It has
also been successfully used to deepen understanding of reasons for non-attendance and hostility towards
the arts. Again, the purpose of this type of research is normally to enable better market segmentation and
improved targeting of potential audiences. For example, different factors motivate visitors to attend a
gallery specifically for a special event or exhibition, compared to visiting the gallery’s permanent col-
lection (Axelsen, 2006).
At its simplest, motivation research involves asking direct questions of visitors as to why they chose a
particular event in a particular venue on a particular day and time. But since many people are unable to
describe exactly why they act as they do, some motivation research involves a variety of less conven-
tional techniques in seeking a deeper understanding of the more covert or even unconscious reasons for
attendance.
Case Study 3.1 Web motivations: Indianapolis Museum of Art

Visits to a museum’s website can be better explained in relation to the motivations of visitors than
by their demographics. When Fantoni et al. (2012) asked visitors to the Indianapolis Museum of
Art (IMA) website about the reasons for their visit, they identified five visitor typologies based on
their motivations:
Planners: Planning a visit to the museum (50 per cent)

Researchers: Finding specific information for research or professional purposes (16 per cent)
Hobbyists: Finding specific information for personal interest (21 per cent)
Browsers: Engaging in casual browsing without looking for something specific (10 per cent)
Buyers: Completing a transaction on the website (3 per cent)
These motivations were found to relate to different website usage patterns. For example:
The average number of pages visited ranged from 6.8 for the Planners, to 15.4 for the Buyers
26 per cent of Researchers downloaded content, compared with 8.4 per cent of Buyers
76.5 per cent of Browsers were new visitors, but that was true of only 56.1 per cent of Buyers
Those least likely to be USA-based were the Researchers – 82 per cent compared with
99 per cent of Planners
Average time spent on each page was 57 seconds for researchers, but only 45 seconds for
hobbyists
Motivations were also found to influence the type of content accessed. For example, the web
page about visiting the museum was accessed by Planners, Hobbyists and Browsers, but not
Researchers or Buyers.
Sources of traffic also varied in relation to website users’ motivations. Those most likely to visit
the site deliberately, rather than stumbling across it, were the Researchers and Buyers. Referral
traffic – those who arrived at the IMA website via links from other websites – was highest among
Hobbyists, Researchers and Browsers, but lower for Planners, suggesting that there was room for
improvement in the way that IMA was promoting itself on local or tourism websites.
Sources: Culture24, 2014; Fantoni et al., 2012; Falk, 2009.
Competitor research
It is often argued that arts events do not face direct competition, as it is difficult to imagine two venues
within close geographical proximity of each other offering identical programmes at the same time.
Indeed, in some cities, cultural organizations use an ‘anti-clash’ diary to ensure that this doesn’t
happen.
Nonetheless, as discussed in Chapter 9, audiences do have choices as to how, when and where to spend
their money, and different venues and programmes will be competing for a share of that money over a
longer period of time. For this reason it is important for arts organizations to understand how their
audiences perceive them in comparison with other similar organizations. Competitor research can help
organizations to understand these perceptions and then to differentiate themselves positively from other
providers of similar arts services. Failing to do so can lead to an ill-defined image, potential confusion,
and consequent rejection by key target audiences.
Product research
This is quite a difficult area in the arts. Commercial and industrial organizations conduct product research
to help them improve the products and services they offer to their customers and to identify demand for
new developments. This reflects the overall objectives of most firms in the private sector, which are related
to profitable trading activity. If demand exists which can be supplied, then there is an opportunity to make
money. Perhaps the nearest the arts come to this kind of research is when cinema producers try out
different endings on preview audiences, or television companies research the popularity of particular
characters in long-running soap operas. Quite apart from the question of aesthetic integrity, the expense
and complexity of such a procedure renders it out of the question for the vast majority of arts organizations.
Product research in the more tangible areas of ancillary products, such as retailing and catering, is more
viable, of course. This type of research can help to identify both inadequacies in existing provision and
demand for new facilities and services which may improve audience perceptions of a venue, encourage
new attenders and consolidate customer loyalty.
Pricing research
For commercial arts organizations, this can help in the setting of entrance fees or ticket prices. Revenue
from the box office or entrance fees can be maximized if the organization has done some pricing research
to help it understand the monetary value that audiences will place on the experience they are expecting
to enjoy (there is more discussion on this in Chapter 6). In the non-profit sector, pricing research
can also be used to help the formulation of pricing policies that will promote wider access to the arts.
Box 3.2 Measuring price sensitivity

The Van Westendorp Price Sensitivity Meter was developed as a method for assessing how much
consumers are willing to pay for a product. This method asks customers four basic questions to
assess their willingness to pay for a ticket, service or experience.
At what price would you consider the product to be so expensive that you would not consider
buying it? (Too expensive)
At what price would you consider the product to be priced so low that you would feel the
quality couldn’t be very good? (Too cheap)
At what price would you consider the product as starting to get expensive, so that it is not out of
the question, but you would have to give some thought to buying it? (Expensive/High Side)
At what price would you consider the product to be a bargain – a great buy for the money?
(Cheap/Good Value)
By combining and graphing the responses of a whole group of people to all four questions, it is
theoretically possible to identify an optimal price, and the upper and lower limits of the ideal
price range.
But using this type of research for setting prices is not without its critics, who suggest that not every
product has a price below which it is so cheap that its quality would be in question. In some cases,
the lower the price, the more people will buy. Also, for some new arts and cultural experiences,
such as a new play or an exhibition by an unfamiliar artist, audiences and visitors may have little
information or comparators on which to base their evaluations.
Source: Isaacson, 2012.
An understanding of the key influences on audience price sensitivity can be gained by experimenting
with different price levels and monitoring the associated attendance figures.
Promotional research
This is normally undertaken to assess the effectiveness of different media, messages and promotional
techniques in attracting audiences. Especially if money has been spent promoting an event or a season, it
is important to be able to gauge the cost-effectiveness of the chosen methods and media of promotion.
Promotional research attempts to gather information which will identify the most persuasive promotional
techniques, normally by linking them to attendance figures. Under some circumstances it may also be
useful to pre-test advertising campaigns, to identify the most effective visual or verbal creative concept,
and to identify the target audiences to whom a particular campaign appeals.
Website and social media analytics can provide valuable insight into the effectiveness of an organ-
ization’s email and digital campaigns. Deciding how well a particular channel is working and whether the
investment in time and resources is delivering positive returns is a valuable exercise: social media are not
cost-free, even if the publishing platforms and channels are.
Policy research
National funding bodies use marketing research to help them make recommendations about the levels of
arts provision and the allocation of resources. Information about national and regional public attitudes
towards the arts, as well as attendance and participation figures, is invaluable in creating a strong case for
public funding of the arts. Research can demonstrate economic benefits from the arts, such as spending in
restaurants or attraction of tourists. It could indicate the effect of arts facilities on the image of a town, or
on local or national pride. Marketing research can also be used to detect audience trends and attitudes in
other countries which may have domestic implications or simply provide early warning of likely
developments at home.
Case Study 3.2 Taking Part in culture

Taking Part is a major survey of cultural and sports participation in England. Run by the
government’s Department for Culture Media & Sport since 2005, it provides the country’s
recognized official statistics on arts and cultural engagement among adults and children.
The survey findings are used to assess the effectiveness of public spending on culture and sport, as
well as to inform the development and impact of government policy. More specifically, it aims to:
Provide a central, reliable evidence source that can be used to analyse cultural and sporting
engagement, providing a clear picture of why people do or do not engage
Meet the needs and interests of other individuals and organizations who use Taking Part data
Underpin further research into the drivers of engagement with cultural activity, and the value
and benefits of this
Information needs
Every year the survey collects information about involvement in a wide variety of arts events, and
attendance at museums and galleries. It also asks about motivations and barriers to engagement,
including information on frequency of participation, reasons for participating, barriers to
participation and attitudes to these activities. By gathering information on demographics, such
as age and education, and related areas including social capital, activities while growing up,
volunteering, charitable giving, and TVand internet use, the survey can paint a picture that reveals
not just how much public engagement takes place in England, but also by whom.
Methodology
Taking Part is a random probability survey taken from a nationally representative sample of adults
aged 16+ and children aged 5–15. In 2015–2016, over 10,000 adults and 684 children were
interviewed over a 12-month period in their homes by trained interviewers using Computer
Assisted Personal Interviewing (CAPI).
Some of the survey respondents are re-interviewed annually to enable the research to reveal
change over time at the individual level. These people are asked reasons for any changes in their
participation, especially in relation to major life events. This longitudinal element helps to build a
picture of how changes in circumstances and other life events can help or hinder participation and
engagement, and for how long.
Questions
The Taking Part questionnaire asks respondents whether they have attended or participated in a
range of arts activities, and these two forms of activity are combined to provide an overall measure
of arts engagement. The definitions of ‘arts’ is wide-ranging. Activities covered in the research
range from attending events such as live music performances and seeing a play/drama, to joining a
book club, practising circus skills and painting.
Digital engagement with the arts is a particularly wide category, defined as viewing or
downloading a performance or exhibition, discussing the arts or sharing art that others have
created, uploading or sharing art that you have created yourself or finding out how to take part or
improve your creative skills. For museums and galleries it is defined as using websites to look at
items from a collection, find out about a particular subject, take a virtual tour of a museum or
gallery or view or download an event or exhibition.
Findings 2016
The survey findings for the year 2015–2016 were disappointing for the arts sector, but more
comforting for museums and galleries.
Despite decades of public investment aimed at increasing and diversifying engagement with arts
and cultural activities, 76.1 per cent of adults had engaged with the arts at least once in the last
12 months – a proportion almost identical to 2005–2006 (76.3 per cent) and a significant decrease
on the 2012–2013 high of 78.4 per cent. 61 per cent of adults had engaged with the arts three
or more times in the previous 12 months, which was also a statistically significant decrease from
62.5 per cent in 2005–2006. The trend for museums was more positive. 52.5 per cent of adults
visited a museum or gallery, a significant increase in attendance since 2005–2006 (42.3 per cent).
General engagement with cultural activity among adults from lower socio-economic groups, more
deprived areas of the country and minority groups (including black, Asian and minority ethnic groups
and those with a disability) remains stubbornly lower than among those from higher socio-economic
backgrounds, less deprived areas and more common demographic groups, though the gaps in
engagement between more and less deprived areas of the country are narrowing. The gap in arts
engagement between adults living in the most deprived and the least deprived areas of the country has
decreased from 26.8 percentage points in 2009–2010 to 14.6 percentage points in 2015–2016.
Sources: TNS BMRB, 2016; Department for Culture, Media & Sport, 2016.
Conducting marketing research

If marketing research is to be an effective tool for improving the quality of marketing decisions, it needs to
address three key questions systematically:
What do we want to know?

The first stage of marketing research should be to examine the marketing problem or opportunity which
requires further investigation and this should lead to a set of specific objectives for the research. Having
set clear, unambiguous objectives it is possible to pinpoint the nature of the information that must be
obtained to help solve the problem or develop the opportunity.
How can the information be obtained?

Secondly, it is essential to identify who can provide the information and the best method for collecting it.
The most visible aspect of marketing research is the collecting of information using surveys. While
surveys are very important, they are but one mechanism for finding out about people, and may be neither
the most appropriate nor the most cost-effective method of data collection.
How will we use the findings?

Marketing research provides facts, but information is created when the facts are interpreted in the context
of the original problem. The final stage, therefore, is for managers to make sense of the findings and
ensure that they influence decision-making in the organization. Without intelligent interpretation,
research is at best worthless and at worst can be misleading and dangerous.
The rest of this chapter explores the techniques and research methods that will help managers to answer
these three questions for themselves in their own organizations.
Research objectives
Why set objectives?

Research objectives are explicit statements of what the organization wants to know. They are important
for two main reasons. Firstly, they are a constant reminder to managers of what they are trying to find out.
This can prevent time and energy being wasted on the collection of information which will not ultimately
be of use in solving the marketing problem facing the organization. Secondly, they can provide a
benchmark or target against which the results of the research can be measured. This enables managers to
assess whether the research was effective, which can be a crucial activity for arts organizations, par-
ticularly in the subsidized sector. If money is scarce, marketing research may be viewed as a luxury and
marketing managers are likely to be required to justify their expenditure in this area.
Categories of objectives
Broadly speaking, research objectives can be divided into three categories: exploratory, descriptive and
causal.
Exploratory objectives tend to be quite broad in scope and are normally specified when an organ-
ization feels that it needs to understand more about certain marketing issues. Academic research into
cultural activity is often exploratory, attempting to understand what lies behind the attitudes and
motivations of arts attenders, leading them to behave in the ways they do.
Descriptive objectives are usually set when an organization needs more concrete evidence to support
specific marketing decisions. Audience research reports are normally descriptive. They describe an
audience or potential audience by their characteristics and preferences, so that the relationships between
different characteristics and preferences can be examined. For example, it would be possible to design a
survey which investigates the age, gender and socio-economic profile of a regional opera audience,
as well as their musical preferences. It could then be determined whether operagoers are also interested in
orchestral concerts, pop concerts or ballet, and whether any particular age group, social class or gender is
more likely to prefer one of those art forms to another. This type of information can be very helpful in
targeting new audiences.
Causal objectives are set with a view to identifying cause-and-effect relationships, in an attempt to
explain why things happen. Experiments are widely conducted to test alternative prices or concessions by
monitoring their impact on audience size and composition. This type of research could be used for
developing access policies, or simply for assessing audience price sensitivity. Similar exercises can be
performed to assess the impact of different advertising media, programme design or even interval length,
thus providing information to help with promotion and programming decisions.
In practice, a major marketing research project may involve all three types of objectives. Exploratory
research could lead to the design of a survey which would enable an audience to be defined and
described. The significant relationships between the different characteristics and preferences of this
audience could then be further investigated in a causal study. Nonetheless, the distinction between
these types of objectives is very useful in helping managers to focus systematically on the purposes of
the research.
Research methodologies
Process of research
Objectives determine the process by which marketing data should be collected. Generating marketing
information requires the collection of marketing data, the term ‘data’ referring to the facts and figures that
must be gathered from either secondary or primary sources to achieve the research objectives.
Secondary research, sometimes referred to as desk research, involves the gathering together and
analysis of relevant data that exist prior to the start of a marketing research programme. The researcher is a
secondary user of already existing data, hence the name. Internal secondary data already exist within the
organization conducting the research; external data have been collected outside the organization, for
example by the government, by funding bodies or by commercial marketing research houses. Even
though these types of data were originally collected for a purpose other than marketing research, they can
still be very valuable to the researcher, even to the extent that the more costly processes of collecting
primary data can be avoided.
Primary research involves generating original data. The organization determines exactly what
information is necessary and from whom, and then sets about acquiring it. The data are thus specific to the
purpose for which they have been acquired. Primary research often takes place after secondary research,
to provide a more complete set of answers to a researcher’s questions.
Type of research
Having set research objectives, it is also possible to identify the type of study that will be most
appropriate.
Qualitative research can reveal why people act as they do, or how they think and feel about the
experiences that the arts are offering them. It usually involves interviews with small numbers of people,
Box 3.3 Eavesdropping at Tate Modern

A project at Tate Modern set out to explore people’s reactions around art – a process that could help
to determine how best to conduct visitor research.
In conventionally hung exhibitions of classical paintings, people tend to move silently from one
painting to another, pausing only to read the information labels. In this situation, ‘dwell time’
around an exhibit can be an effective measure to indicate the public’s appreciation of a work of art.
But by observing and recording the conversations of groups of gallery attenders gathering
around These Associations, Tino Sehgal’s commission for the Turbine Hall, researchers found
that this approach to assessing visitor opinions is not universally valid. Visitors formed their
opinions in very different ways, and on the rare occasions when they commented “It’s beautiful”
or “I like it”, these expressions were always within the context of ongoing social interaction.
Their conversations revealed what These Associations meant to them in that specific situation,
suggesting that simple linear measures of ‘like’ or ‘dislike’ are inadequate for understanding the
visitor experience of a work of art.
Source: Albert, 2016.
though other techniques such as observation and experimentation may also be useful in getting to the root
of audience behaviour. In the arts, qualitative research is particularly useful as it is able to explore the
subtleties of people’s reactions to the aesthetic experience. It may seek to explore issues such as:
Motivations and inhibitions for participating in the arts

What people are looking for from the arts
Perceptions of different art forms
Reactions to specific productions, titles and artists
Reasons for success or failure of productions or events
Appropriate types of promotion
Sources of influence over audiences (reviews, advertising, word of mouth)
The perceptions of sponsoring organizations
Quantitative research generates information about how many people hold certain views or fit into
certain categories. Conducting this type of research often involves among a sample of the population of
interest to the researcher. The findings can then be interpreted with the help of statistical techniques, and
assumptions can be made about the whole population from the information generated among the sample.
This is a very common form of research in the arts, and is a popular method for investigating the nature of
audiences.
Duration of research
The necessary duration of research is also implied by the research objectives.
Continuous research examines an issue or problem on a regular basis in order to monitor changes that
are occurring over a period of time. For example, funding bodies are interested in changes in arts
engagement patterns. England’s Taking Part survey provides a good example of this (see Case Study 3.2).
Ad hoc research is the term used to describe a one-off piece of research to obtain information relating
to a particular issue or problem. The findings generated by this type of research tend to be reported in such
a way as to help a specific marketing decision to be made.
Sources of secondary data
Internal sources
Arts organizations hold a lot of useful marketing data by virtue of the activities they perform as part of
their core activities. Those that charge for admissions are in a position to collect vast quantities of data
about attenders and their transactions through their box offices and their online ticket sales. Those that
don’t charge, but perhaps live-stream their content or use their website to share additional resources, such
as for educational purposes, can gain important insights into their effectiveness through their website
analytics. And if social media are used to broaden or deeper relationships with audiences, the data gained
from this can help them build better relationships in the future.
Analytics
Websites, social media, digital advertising and apps (discussed in Chapter 7) all generate statistics that can
reveal customer behaviour and potentially trigger a marketing response. For example, analytics can be used
to find out how many people open an organization’s emails; which of its web content is most viewed; which
search terms bring most traffic to its website; and even the type of web browser or mobile device visitors use.
Website analytics can play an important role in helping marketers improve the experience of their
online visitors. Deciding what to measure is crucial. Some visitor analytics are worth looking at regularly,
to enable comparisons and understand trends. These might include:
Page views, downloads and visitor numbers for different articles, collection items, videos and
updates, to understand the relevance of web content
Time spent on different pages, to understand how compelling content is
Referrers (how people arrived at web content), to see how well distribution efforts are working
Spikes in visitor numbers, to see if they present marketing opportunities
People bouncing back and forth between areas of the site or repeatedly using a site search to identify
site navigation problems
Geographic location of visitors, to locate audiences and potential audiences
The language visitors use in their web browsers, to see if translations of the visitor information pages
are needed
Google Analytics is the most widely used analytics measurement tool on the internet and can provide
much of this headline data free of charge. But for addressing specific marketing objectives, actionable
analytics can be more useful. These are measures that tie actions by website users to observed results. A
cohort study can examine the behaviour of groups of website users who share common characteristics or
experiences, within a defined timescale. For example, to evaluate the effectiveness of video for attracting
potential attenders, a gallery or museum might like to investigate how many web visitors who viewed its
video in the period before a new exhibition actually bought tickets. So-called ‘vanity metrics’, including
web hits or number of downloads, can document the current status of a website, but they contribute little
to an understanding of how this came about or what to do next. Without actionable analytics, the
information that is being presented may not always have a practical application.
Social media analytics can provide marketers with useful information to help them direct their efforts
and resources most effectively. Fontein (2016) proposes five dimensions where they can be useful:
Presence: the number of people engaging with an organization or its products and services on social
media will indicate levels of awareness. The number of ‘mentions’ provides a useful measure of
overall reach.
Box 3.4 Key performance indicators

Tasich and Villaespesa (2012) suggest that, before deciding which analytics to use, organizations
should identify key performance indicators and ask themselves a number of questions:
Why should this performance indicator be measured?

What is the target?
What actions will be taken to influence the achievement of the performance indicator?
Which analytics tools are needed to report it?
Whose responsibility is it to achieve the performance level?
Category Key performance indicator
Reach and brand Visits and searches

Audience Relevant target segments (e.g. international visitors)
Conversion Revenue from e-commerce activities
Social engagement Pages per visit, time on site, repeat visitors, user-generated
content, click-throughs
Usability Satisfied users (comments, online survey), ease of navigation (user testing),
accessible content
Technical Load time, broken links, browser compatibility
Figure 3.1 Key performance indicators (adapted from Tasich & Villaespesa, 2012).
Engagement: the number of retweets, social shares, comments and referral traffic to a website from
social media are all measures that indicate depth of engagement and indicate the power of social
media tactics.
Influence: measuring ‘share of voice’ enables an organization to understand how its social media
profile compares with its competitors. The number of social media ‘influencers’ – users who
themselves have large numbers of followers – who engage with an organization is another good
indicator of success.
Action: tracking the actions people take as a result of engaging with social media enables an
organization to assess the extent to which its investment in social media is generating returns.
Typical ‘conversions’ – actions that ultimately lead to a sale – may be financial transactions for
tickets, but could equally be newsletter sign-ups, content downloads or other outcomes.
Internal measurement: monitoring where the organization devotes time and energy to social media
can improve efficiency. The number of social media posts per day indicates how consistently the
organization is engaging with its audience. Finding the optimum number of posts tends to involve
some trial and error, but is an important aim; too little presence is a lost opportunity, and too much
exposure can lead potential audiences to switch off altogether.
Ticketing systems
Arts organizations that use integrated customer relationship management (CRM) software in the form of
a box office ticketing system are in a good position to use that data to generate a better understanding of
their audiences, which in turn can inform their marketing decisions.
A ticketing system is a type of CRM system that links ticketing transactions to a single central cus-
tomer database. The database at the heart of a good ticketing system contains all the information on
customers and their interactions with an organization, from every ‘touch point’. Whenever a customer
buys a ticket – whether online, over the phone or in person – the details of that transaction can be stored
and linked to their personal data, other transactions at the venue, and information from their membership,
subscriptions, donations, and even their answers to survey questions. Electronic point of sale (known as
EPOS) systems such as touchscreen tills and tablets in cafes and shops can enable data about expenditure
on food, drink and other purchases to be linked to customers’ records. Such is the level of data integration
possible that some venues – especially those with high levels of online booking – get rid of their box
offices altogether and sell tickets over the bar or even out in the street.
A wide range of useful data can be gathered as part of processing the ticket sale, provided that the box
office system and any linked online ticketing is configured to collect it. Name and address, ticket price,
event attended and time of performance are obvious box office data fields, but Tomlinson, Roberts and
Allpress (2006) offer useful advice on some of the details that should be considered to make the most of
this source of secondary data. For example:
Name: first names don’t always indicate gender, so include a field for this in case you want to want to
track the profile of your audience, or use the information in a personalized direct marketing
campaign.
Address: this could be a home address, business address or temporary address, for example, a
holiday home, so ensure that it is possible to make a note of which applies. The extent of a venue’s
catchment can be determined from postcode data, and for some venues it’s important to understand
what proportion of their sales are to local people, as opposed to commuters or tourists.
Date of birth: this can be useful for profiling different audiences, and can potentially be used to send
personalized birthday cards.
Events/performances attended: categorizing events enables organizations to analyse audience pat-
terns for different types of performances and use this for marketing and programming purposes. This
is not always straightforward. For example, decisions have to be made as to whether ‘opera’ should
have a category of its own, or be tagged as a ‘musical theatre’ event, or even simply ‘music’.
Level of engagement: some customers are more than simply ticket buyers and may have multiple
relationships with the organization, for example as a donor, a sponsor, a Friend, a subscriber, a volunteer
or a board member. This type of data can be particularly useful for targeting specific marketing
tactics, such as preferential booking terms, and for identifying prospects for fundraising campaigns.
Time of booking: for certain types of performances, attenders are more likely to buy tickets on
impulse, while for others they may book months in advance. Understanding booking patterns is key
to delivering timely marketing messages, and ensuring that discounting to encourage sales isn’t
triggered too early.
Type of ticket: if price discounts are offered, it’s important to identify the reason for the concession.
It’s not very helpful to know that a person claimed a £5 discount unless it’s possible to tell whether
this is because they are a student or a senior citizen, or redeemed a discount voucher.
Box office data can help marketers answer both strategic and tactical questions. For example, which
customers are most likely to attend contemporary dance productions? What price concessions will be
most effective on which days of the week? At what point should discounts be introduced to improve
ticket sales for a low-selling event? How far are audiences willing to travel to reach the venue? When is
the best time to start an email campaign for a new production?
Historic booking patterns can inform forecasts. For example, how far in advance did people book? Did
they come in groups or alone? Which were the most popular seats? What was the average ticket price?
Case Study 3.3 Christmas bookers

Autumn is an important time for theatres around the UK to increase their marketing activity in the
build up to the Christmas period, in preparation for either a traditional pantomime, large-scale
musical or other seasonal offering.
Ticket sales increase each month in the lead-up to Christmas, but this isn’t the end of the story.
By examining the sales patterns of more than 50 venues across the UK, data agency Purple Seven
was able to describe a national trend which suggests that a venue can expect to sell as many tickets
to customers in November as it did in the whole year up to and including September. In fact the only
period that generates more sales is after the show has opened, which accounts for 32 per cent of all
individual tickets (excluding agents and group bookers).
Examining exactly who books these tickets is revealing. Existing customers – those who have
previously bought tickets at the venue – are more likely than new customers to make a booking
earlier than September. They are likely to already have received marketing communication from
the venue, such as a season brochure, and be booking off the back of a good experience during the
previous year.
The trend for new customers is to purchase closer to the event. In March, just over 10 per cent of
bookers for the Christmas show are new. This is compared with November, when 37 per cent are
new bookers, and December when 45 per cent of bookers are new.
November and December account for 60 per cent of all bookings from a venue’s new customers,
making this the best time to ensure that marketing communications appeal to a wider audience –
not just those who have attended before.
But communicating with a ‘potential’ audience is a lot more difficult than communicating with
existing bookers whose records are held in a box office system. To overcome that problem Purple
Seven developed a proprietary software solution, Arts Dimensions, to help its clients find their
‘untapped potential’ market.
But the story doesn’t end there. It’s important for venues to keep engaging with their new
audiences after the Christmas show is over. On average, only one in ten new bookers will return to a
25%
23% 23%
14%
7% 7%
Up to and October November December December January

including (pre-opening) (post-opening)
September
Figure 3.2 Pre- and post-opening statistics.
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
October
November
December
January
February
March
April
May
June
July
August
September
October
November
December
2012 2012 2012 2013 2013 2013 2013 2013 2013 2013 2013 2013 2013 2013 2013
Existing Booker New Booker
Figure 3.3 2012–2013 graph.
venue. But new bookers who attend a Christmas show buck this trend. Analysis at the 50 venues
found that 38 per cent of their new audiences made a booking at the same venue within the next ten
months. So all in all, giving audiences a good experience at a Christmas show is vital: it’s a great
vehicle for cementing long-term relationships with new customers.
Source: Mitchell, 2014a.
Current patterns which depart significantly from forecast patterns can be flagged up and remedial action
taken. If advance ticket sales are slower than expected, more effort can be put into promotion, or price
promotions introduced; but if an event is selling more quickly than expected, plans can be drawn up for
switch-selling disappointed customers. There may also be opportunities to manage demand through
dynamic pricing (see Chapter 6).
A ticketing system can also provide a statistical overview of the audience for a particular venue. This
information is particularly useful for touring companies trying to identify towns and cities where the
profile of local attenders is similar to the profile of attenders for their own art form. Catchment areas,
customer demographics and attendance patterns can all be helpful.
External sources
Research published by other arts organizations, academic institutions, research agencies and gov-
ernments may be able to shed light on a particular research objective. This type of secondary data
is often available online, and often free of charge, so it is normally worth initially trawling the
internet to see if a particular marketing issue has been explored before, rather than jumping into
primary research and potentially reinventing a wheel. It is important, though, to check who the
originator of the research was, and understand their reasons for doing it. This will indicate whether a
particular slant has been taken in the interpretation of the findings. For quantitative research, the
nature and size of the sample is very relevant. Findings should be treated with caution if based on a
small number of respondents, particularly if comparisons are made between different groups of
people.
Commercial research
Although not specifically related to an arts organization’s own audiences, surveys conducted by
others for other purposes can sometimes offer important insight into their likely characteristics and
behaviour.
For example, the TGI (Target Group Index) is a commercial marketing and media survey that collects
regular information from a representative sample of adults – around 25,000 in the UK, and a total of over
700,000 in more than 60 countries across the world. This continuous survey asks questions covering
consumer purchases, attitudes, motivations, media and digital consumption, leisure activities and
demographics. It includes questions on frequency of attendance for nine types of art forms, namely plays,
opera, ballet, contemporary dance, classical music, jazz, art galleries or exhibitions, cinema and popular
or rock concerts. The data it collects about a very wide range of other consumer purchases can be related
to the theatre attendance data to help arts marketers understand more about the other types of products and
services that arts attenders use. Understanding more about their exposure to different media, for example,
can help art marketers plan more targeted promotional campaigns.
Geodemographic profiling systems

While box office information allows marketers to confirm the intuitive pictures they form about their
audiences and respond accordingly, linking this to external data based on geodemographic profiling can
add powerful insights to guide marketing activity.
Geodemographic profiling works on the assumption that people living in similar neighbourhoods (in
the UK defined by postcodes) are likely to have similar interests and purchasing habits. They offer arts
marketers a way of identifying areas within their catchment where residents are likely to share similar
attitudes or behaviour patterns in relation to the arts and culture, and target the groups that are most likely
to attend. (see Chapter 2). These systems draw on geographic, demographic and lifestyle information to
link the places where people live with their underlying characteristics and behaviour. Some of them
describe generic population behaviour patterns, built from public data and large commercial databases.
Others are arts sector-specific, built from data that indicate how populations engage (or do not engage)
with the arts.
One generic system, Acorn, classifies UK postcodes into 62 ‘types’ that give a distinctive picture of the
likely aspirations, attitudes and general purchasing behaviour of residents in each area (CACI, 2016a).
Each type is given a name that aims to describe the people who live in that postcode: ‘Wealthy
Countryside Commuters’; ‘Smaller Houses and Starter Homes’; and ‘Career Driven Young Families’ are
all examples. Nationally, certain types are more likely to engage with the arts than others, and different
types will engage in different ways. Type 12, ‘Retired Empty Nesters’, for example, are described as “less
likely than average to use the internet”, “less likely to use social media”, and “not particularly keen on
online shopping” but “leisure time might involve gardening, wildlife, or arts and crafts”. Type 13,
‘Upmarket Downsizers’, may also be retired, but are more likely to use digital technologies and name
‘cultural events’ among their interests (CACI, 2016b).
Another system, Mosaic, works along similar principles but uses particularly colourful labels to
describe its 66 ‘groups’. Group 8, ‘Bank of Mum and Dad’, refers to “Well-off families in upmarket
suburban homes where grown-up children benefit from continued financial support”; while group 22,
‘Boomerang Boarders’, comprises “Long-term couples with mid-range incomes whose adult children
have returned to the shelter of the family home” (Experian Limited, 2014).
A cultural sector-specific system used in England is The Audience Agency’s ‘Audience Spectrum’,
which classifies patterns of arts consumption and attitudes towards culture, how people spend their
leisure time and what competes with the arts for people’s attention. Audience Spectrum combines data on
cultural engagement (from the national Taking Part survey, see Case Study 3.2) with wider lifestyle
context and decision making, demographic profiles, socio-economic status and relative access to cultural
opportunities.
Ten distinct consumer segments are identified among English adults, some of them highly engaged
with the arts, some who have some engagement, and others that aren’t engaged at all. Pen portraits
describe the segments; the proportion of English households that fall into each segment is estimated
(Figure 3.4).
Detailed descriptions reveal the likely characteristics of each segment. For example, the ‘Trips and
Treats’ segment – the largest group in England – consists of people who are comfortably off and living in
Metroculturals
Prosperous, liberal, urbanites interested in a very wide cultural 5%
spectrum
Commuterland Culturebuffs
11%
Affuent and professional consumers of culture
Experience Seekers
Highly active, diverse, social and ambitious, engaging with arts on a 8%
regular basis
Dormitory Dependables
Suburban and small towns interest in heritage activities and 15%
mainstream arts
Trips and Treats

Mainstream arts and popular culture influenced by children, family 17%
and friends
Home and Heritage

10%
Rural areas and small town, daytime activities and historical events
Up Our Street
Modest in habits and means, popular arts, entertainment and 8%
museums
Facebook Families
Younger suburban and semi-urban, live music, eating out and 12%
pantomime
Kaleidoscope Creativity
Mixed age, low level engagement, free local events, outdoor arts 9%
and festivals
Heydays
Older, less engaged, crafts, knitting, painting, sheltered housing, 5%
church group or community library
Figure 3.4 Audience Spectrum (from The Audience Agency, n.d.).
the heart of suburbia. They are led by their children’s interests and influenced by friends and family. For
them, arts and culture is not a passion, but they are reasonably culturally active, as well as having a wide
range of other leisure interests. They tend to have a strong preference for mainstream arts and popular
culture like musicals and familiar drama, mixed in with days out to museums and heritage sites.
For marketing purposes, geodemographic information can be very valuable. Arts organizations can
tag their own database with geodemographic profile codes to help them understand the profile of their
current audiences. Although these systems can’t predict exactly how each individual in a given segment
will behave, marketers can assess how representative their customers are of the local population and
identify customer groups that offer potential for increased sales. Promotional activity can then be targeted
more accurately. For example, the ‘Trips and Treats’ segment described above could be targeted with
multi-ticket offers to encourage them to attend more often, using offers such as ‘4 shows for the price of 3’
combining mainstream events with more adventurous ones to reduce the financial risk of trying out
something different.
Box 3.5 Finding flourishing families

Venue Cymru in Llandudno compared the Acorn profile of its current attenders with the profile of
the population living within its 90-minute drive-time catchment. Certain types of attenders on its
database were found to be over-represented compared with the local population, while others were
under-represented (Figure 3.5).
The most important types are those that make up a large share of the population. While the venue
had been successful at attracting ‘Affluent Greys’ and ‘Settled Suburbia’, it had been less
successful among ‘Struggling Families’ and ‘Blue Collar Roots’. However, arts attendance among
these groups tends to be fairly low, so this was not of great concern. Of the types in the middle
column, ‘Secure Families, ‘Wealthy Executives’ and ‘Flourishing Families’ offered the best
potential for attracting new audiences, as they all have disposable income and arts attendance tends
to be fairly high among them.
‘Flourishing Families’ often take family to the cinema, tend to be in managerial and professional
occupations, and most have two cars. Postcode mapping enabled Venue Cymru to identify
Anglesey, Bangor, Prestatyn, Chester, Northwich and Wrexham as fertile locations for a
promotional campaign to this group.
Source: Humphreys, n.d.
Prevalence among venue attenders
Prevalence in High Average Low

local population High Affluent Greys Secure Families Struggling Families
Settled Suburbia Wealthy Executives Blue Collar Roots
Flourishing Families
Average Prudent Prospering Pro- Burdened Singles
Pensioners fessionals Post-industrial Families
Starting Out
Low Aspiring Singles High-rise Hardship
Educated Asian Communities
Urbanites Inner-city Adversity
Figure 3.5 Prevalence in local population (from Humphreys, n.d.).
Collecting primary data

Audience or visitor surveys are very useful and effective ways of collecting certain types of primary data
and the final part of this chapter is dedicated to the planning and implementation of such surveys.
However, they are not the only methods of collecting primary data. Neither are they necessarily the best
methods for collecting certain types of data, in particular qualitative data. Decisions about the best
method or methods to use for collecting primary data should always be made by considering first of all the
objectives of the research.
Four methods are commonly used to collect primary data:
Observation
Experimentation
Interviews
Survey
Contextual observation
This is a technique that involves watching people, their behaviour and their actions as they use a product
or service. It is particularly useful when people are unable to give accurate verbal accounts of their
behaviour, perhaps because they cannot remember sufficient details of it. In the arts, it is a technique that
can shed light on the ‘user experience’, for example if you wished to know what path an individual had
taken around an exhibition, which exhibits he or she stopped to look at and for how long, or how
audiences spend the interval during a performance. Observation may be overt, whereby the subject of
research is aware of the observation taking place, but commonly the observation is covert, so that the true
relationship between the audience and an art form can be studied without interference from the
researcher.
Galleries and museums have benefited greatly from this type of research, and technology is making it
easier to gather this type of data. Bluetooth Low Energy (BLE) beacons provide an effective means of
gathering data to gain insights into museum visitor behaviour and using this to optimize exhibit locations
and museum layouts. A person’s entire visit can be tracked, including dwell times at certain exhibits and
time spent at different locations within the museum. This makes it possible to find out whether an exhibit
itself is popular, or its popularity is related to its placement at a particular location.
Experimentation
Primary data can be gathered by experimentation, which is a particularly useful technique when a
decision is under consideration and the outcome cannot be predicted either from existing experience of
the organization itself or from the experience of other organizations. In an experiment, the researcher tries
out some marketing activity on a small scale, observing and measuring the results, while controlling as far
as possible the effects of factors other than the marketing action being taken. The key advantage of
experimentation is that the researcher chooses which factors (known as variables) are going to be tested.
This enables conclusions to be drawn as to the nature of causal relationships between variables.
Suppose, for example, a theatre wished to assess the price sensitivity of audiences for Shakespearean
productions. Assuming that attendance profiles at previous productions have been compiled, they can be
used as a benchmark (or control) for assessing the impact of a different pricing structure on attendance
profiles. Statistics can be compiled to indicate any significant differences in attendance profiles between
the production in which the experiment is taking place and previous productions, and these differences
can be attributed to the pricing structure. It is important that extraneous variables (i.e. those factors which
Case Study 3.4 The value of digital exhibits

At the Cocoon space in London’s Natural History Museum and at the Treasures Gallery in the V&A
museum, a combination of observation and video-recording using cameras worn by the people being
studied enabled researchers to examine how people use digital exhibits, the group interactions taking
place, and the physical responses to exhibits. The observations led to a number of conclusions:
1. Atmosphere matters
The curved shape of the Cocoon space was described as ‘egg-shaped’ and having a ‘natural’
feel to it, and visitors commented on it ‘feeling modern and inspiring of adventure’. By
contrast the Treasures Gallery incited a sense of exclusiveness and mystery in visitors.
2. Interaction as a shared experience
The size of large touch tables in the Cocoon was crucial to enabling family members to
collaborate and appeared to be successfully enabling families to do something together.
3. Don’t fight the familiar
Children expected the touchscreens on the Cocoon touch tables to react like familiar
technology, such as smartphones and iPads, and interaction was found to be more difficult if
they worked differently. This suggests that unless the effort involved in figuring out how
something works is low, people are likely to wander off.
4. Watching and learning
Visitors observe each other’s behaviour and pick up non-verbal clues about what to do. This
was most noticeable in the V&A’s Digital Dragon interactive, aimed at children, where video
animation was triggered by visitors’ physical actions. This suggests that it may be more
effective and intellectually less threatening for visitors to see staff interacting with technology,
rather than reading text-based instructions.
5. Interaction varies with age
The Digital Dragon interactive allowed multiple people to enter and engage. Younger children
reacted by noticing changes as they happened and tended to concentrate on just the walls
or just the floor. Older children tended to react by seeking connections between actions.
6. Ergonomics
The position and content within dual screens affected concentration. Mixed vertical and
horizontal screens didn’t appear to work if both required visual attention: people concentrated
on one screen or the other.
7. Physicality within experiences
Visitors displayed three types of physical actions and reactions to exhibits, which had strong
social elements:
Explanation – showing others how to do something
Expression – posing or dancing, either alone or with others
Competition – actions intended to outdo other people’s actions
8. Touch
The urge to touch and what it meant to people was a powerful motivation for visitors.
The study noted the general urge to touch things as a discovery aid.
Source: Lewis, 2014.
Box 3.6 Testing times

A small theatre company in San Jose doing a musical production of Babes in Hollywood wanted to
see if they could influence people to buy tickets by using an audience-response video that filmed
people commenting on the show after the opening night and encouraging others to come along. But
would it be worth the time and effort of creating and editing the video?
The company did an A/B test. One email, with just a photo in it, went out to a random 1,800
people on their emailing list, while another 1,800 were sent an email that had the video embedded in
it. In both of the emails was a big ‘buy tickets’ button, and the experiment aimed to find out whether
the people who saw the video would buy more tickets than those who didn’t.
It turned out that, while lots of people watched the preview video, it had no effect whatsoever
on sales.
Source: Evans & Tomlinson, 2013.
are not the subject of the research) are as similar to the control situation as possible. One disadvantage
of experimentation stems from the extent to which the extraneous factors can be controlled. Factors such
as the weather, the economic climate, the promotional campaign, critics’ reviews and the style of pro-
duction may have an equal, if not stronger, impact on audience profiles than the price of a ticket, and if
these vary greatly between the experiment and the control situation, it will not be possible to identify the
true cause of changing attendance patterns.
Interviews
Qualitative data are most commonly collected using loosely structured interviews. Qualitative research
does not demand that a formal structure is imposed on the data collected. Indeed, a formal structure may
so constrain responses to questions which require the articulation of attitudes and motivations that the
findings may be of little real value. The interview is an approach to data collection which is highly
sensitive to the views of the individual being questioned. It attempts to probe beneath the surface of the
responses that can be obtained to a formal question in a survey questionnaire. Two specific techniques are
commonly used, namely depth interviews and focus groups.
Depth interviews are usually prolonged one-to-one interviews, during which the interviewer will ask
questions on a series of topics, but has the freedom to phrase the questions as seems most appropriate and
to order them and probe them according to the responses. The interviewer has only a checklist of points to
cover and will be using mainly open questions to encourage full and explanatory responses from the
interviewee. The interviewer will attempt to move from the general to the specific and from topic to topic
in a logical manner and will also try to monitor the non-verbal behaviour of the respondent. The skills of
the interviewer are of paramount importance in gaining the deepest insight into the behaviour, attitudes
and motivations of the research subject.
Focus groups involve an interviewer, known as a moderator, asking questions of a group of respon-
dents, usually around eight people. They are useful for obtaining several perspectives about the same topic
and gaining insights into people’s shared understandings. Groups provide a social background which
reflects the fact that many of the decisions made by an individual are taken in a group context, particularly
decisions on participation in leisure activities such as the arts. The range of opinions expressed in a focus
group can help individuals articulate their own beliefs, attitudes, opinions and feelings.
The role of the moderator is to act as a catalyst for appropriate conversation between members of the
group, intervening as little as possible except to keep the conversation on the right lines and ensure that all
the required topics are covered. Less assertive individuals may feel group pressure to conform to the
norms of attitudes expressed in the group and dominant individuals can discourage a full discussion of
topics about which they feel strongly. It is the role of the moderator to use group dynamics to draw out
individual beliefs which might not be so freely expressed in a one-to-one interview situation, and to
control excessive behaviours.
The interpretation of the data is usually the responsibility of the interviewer or moderator, who has
observed the data collection first-hand and is therefore in the best position to interpret its meaning.
Interviews are usually recorded then transcribed before the text is analysed, though low-budget studies
may have to rely on notes taken about key issues while listening to the recording. Using the transcript to
code and then categorize interview data is preferable, as this enables large amounts of text to be organized
to discover patterns that might be difficult to otherwise detect. Specialist software can help with the coding,
identifying relationships and counting key words, and word clouds can be used to present the data visually.
As a research approach interviews can stand alone, but as the size of the sample is very small and no
explicit attempt is made to ensure that the sample is representative, the findings from interview research
should not be used to make generalizations about whole audiences or market segments. Interviews do not
provide the statistically significant data generated by the larger samples used in surveys. Used in con-
junction with quantitative surveys, though, this method has few drawbacks. Interviews can be of
immense value in exploratory research or to illuminate further the findings of a survey.
Survey
Research objectives that aim to measure public opinion on an issue require quantitative data and a survey
is likely to be the most appropriate method of data collection. A survey involves identical questions being
asked of a large number of individuals and a systematic record being made of their responses. This
process is usually conducted using a questionnaire. The rest of this chapter will take a closer look at
survey design, implementation, analysis and interpretation.
Conducting a survey
Arts organizations most commonly use surveys in ad hoc studies relating to specific marketing decisions,
and, in particular, to find out more about audiences. Figure 3.6 shows the different stages of a survey.
These stages will now be examined in more detail.
1.
6. 7. 1a.
Analyse the problem
Process Interpret and report Exploratory
and set survey
the data the findings research
objectives
2.
5. 4. 3.
Decide how the
Collect Construct a Identify an
survey will be
the data questionnaire appropriate sample
administered
Figure 3.6 The stages of ad hoc survey research.
Stage 1: Identify the problem to be solved and set objectives for the survey
The researcher should begin by analysing the problem or problems that the research is to address. It is not
enough to have a vague notion that you would like to know more about your audience or believe that you
could attract more visitors if you changed your promotions. Since the purpose of a survey is to get
answers to specific questions, a crucial first step is to state those questions explicitly in the form of
objectives for the survey – when the survey has been completed, precisely what do you want to know? It
may be wise, if not essential, to undertake some exploratory qualitative research at this stage to help
define the questions more clearly.
The next two stages need to be considered in tandem. Decisions have to be made as to who should be
asked to respond to a survey (the sample) and this may affect the way in which it is administered.
Stage 2: Identify an appropriate sample

The individuals chosen to respond to a survey dictate its findings, so it is important for the target audience
for the survey to be clearly defined so that the questions can be asked of appropriate individuals.
Surveys rely upon the statistical fact that the views expressed by a relatively small sample of indi-
viduals are representative of the views of a much larger group, known as the population of interest, if that
sample is chosen in a particular way and is of a particular size. The laws of probability mean that the larger
the sample, the more representative it will be of the population from which it was drawn. A random
sample, in which each person in the population has the same chance of being chosen as any other member,
will be most representative of the views of the whole population. In practice, of course, it can be much
more difficult and expensive to identify a large sample than a small one. Also, the logistics of randomly
selecting a sample can be too complex and expensive for most arts organizations. As a result, rules of
thumb may be more appropriate to guide their sampling than pure statistical techniques.
Three key questions need to be answered when deciding on an approach to sampling:
When should I sample?
The sample needs to take into account variations in audience profile caused by season (e.g. school
holidays), day of the week, time of day, time of year and maybe even weather conditions. It’s important to
recognize that the profile of a museum’s visitors can look very different, for example, during the summer,
especially if the venue is a magnet for international tourism. Also, asking a person about an event they
attended the previous evening can elicit a different response than if they were to be asked a week later
when they have had more chance to reflect on the experience and discuss it with others.
Who should be in my sample?
The goal for sampling is to obtain a good cross-section of responses which will be representative of most
of the audience.
If a questionnaire is administered face-to-face, then an objective selection rule should be applied to
protect the survey from the bias arising from interviewers choosing to approach respondents who make
themselves more accessible and avoid those who are less forthcoming. Possible methods include
approaching, say, ‘every tenth person through the door’, or using a quota system, whereby the inter-
viewer approaches a target number of people in each of several predetermined groups with specific
characteristics, such as age and gender.
If a questionnaire is self-administered, as in the case of email, postal and some in-venue surveys, it
is impossible to impose any tight controls over the composition of the sample. Consequently there can
be serious bias in the survey findings, as those with a higher than average interest in the subject of the
survey – perhaps because they loved an event, or hated it – will take the trouble to complete the ques-
tionnaire. This could mean that, rather than using the findings to draw conclusions about the population as
a whole, it may be prudent to only draw conclusions about sub-sets of respondents who share specific
characteristics.
How many people should be in my sample?
Sometimes it is possible to conduct an exhaustive survey, known as a census, with all visitors or audience
members in the population of interest being asked to complete a questionnaire. This might be possible at a
small participatory workshop, for example. More often, though, a smaller but carefully controlled sample
is more practical. The more questionnaires that are completed, the more likely it is that the responses are
representative of the whole population and not just the individuals who answered the questions. As a rule
of thumb, 500 completed questionnaires should be plenty to provide results which are statistically useful.
Fewer responses can still provide a useful indicator, but can be problematic for analysing sub-groups
within the population. It might be helpful, for example, to understand whether different age groups
respond differently to an issue, or those attending on different days of the week, or those who are regular
as opposed to infrequent attenders, but if the number of respondents in each sub-group is too small the
figures can be misleading.
Box 3.7 Sample sizes

“You don’t have to eat the whole cake to know what it tastes like!” Market researchers use this
analogy to explain the use of samples to represent the views of a whole population. Statistical
techniques enable us to predict just how similar to the population a sample will be. The size of the
sample is critical to accuracy. An estimate of the population’s response can be calculated from the
sample’s response by adding or subtracting the percentages given in Table 3.1. So, from a random
sample of 1,000, if 30 per cent of the respondents give a particular answer to a question, the
percentage agreeing with that response in the whole population would fall somewhere between
27.2 per cent and 32.8 per cent – known as the ‘confidence interval’. In other words, if 30 per cent
of the sample say they have been to an art gallery in the past 12 months, the percentage of the whole
population for whom this would be true would be between 27.2 per cent and 32.8 per cent. But if the
sample size was only 100, then the response for the whole population could fall anywhere between
21 per cent and 39 per cent.
Table 3.1 shows how the responses given by the sample more accurately reflect the views of the
population as sample sizes increase. Note that a sample of 500 can reflect the views of the whole
population with at least twice the accuracy of a sample of 100, but a sample of 1,000 does not
predict twice as accurately as one of 500. There are diminishing returns to ever larger sample sizes.
Table 3.1 Sample sizes
Sample size Response frequency (%)

10 or 90 20 or 80 30 or 70 40 or 60 50
100 5.9 7.8 9.0 9.6 9.8

500 2.6 3.5 4.0 4.2 4.4
1000 1.9 2.5 2.8 3.0 3.1
2000 1.3 1.8 2.0 2.1 2.2
Stage 3: Decide how the survey will be implemented

The way a questionnaire is distributed can influence response rates. Methods of questionnaire admin-
istration fall into five main categories.
Online surveys
The growth of sophisticated proprietary survey software systems coupled with the development of box
office systems over the past decade has brought about a revolution in survey distribution. Mail surveys,
which used to be the mainstay of audience research for ticketed venues, were costly to administer, as was
data analysis. An online survey is a low-cost alternative that enables data to be collected and analysed in-
house. The questionnaire can be longer and more complex than one distributed by other means. It can be
completed at a time to suit the respondent, who can, if necessary, pause halfway thorough and resume later.
Complex routing is straightforward, with the survey leading respondents seamlessly to different sets of
questions, depending on their answer to a previous question. Post-survey data entry is avoided, as responses
are gathered in the online database and keyed in by the respondent. Summary statistics can be viewed while
the survey is in progress, and detailed statistical analysis and charts can be presented at the touch of a button.
There are some disadvantages of this type of survey, the most significant of which is the requirement
for a database of email addresses or other means of encouraging potential respondents to go online to
complete the questionnaire. But for ticketed venues, this is a minor issue as email data captured from
online ticket bookers (but not those who attended with them) can be used for this.
Some venues use online surveys to conduct regular exit surveys after their shows. While the box office
captures behavioural data about the performances and art forms a person attends, their frequency of
Box 3.8 Did you enjoy that?

The online customer feedback survey called ‘Did You Enjoy That?’ triggers audience surveys
straight after a performance, sending bookers a link to a short online questionnaire through which
they can share their reactions to the experiences they have just had. Ten per cent of questionnaires
are completed within the first two hours of customers leaving the auditorium mainly, on
smartphones, and two-thirds complete it within the first 24 hours. By aggregating 200,000 survey
responses to ‘Did You Enjoy That?’ questionnaires, system designers Purple Seven found that
responses represented a sample with a good spread of age groups, including 10 per cent in the 25–
34 category and 25 per cent, the largest proportion, in the 45–54 group. Those most likely to
respond were the 40 per cent of customers who had never been to a venue before.
The findings confirmed that tempting customers back after their first visit is a major challenge for
theatres. 85 per cent of respondents had not rebooked at their original venue during the three years
that followed, and those who rated the event highly were only slightly more likely to rebook than
those who didn’t. The single biggest factor found to influence rebooking was how a new customer
rated a venue. Of all the first-time customers who gave a venue a very low rating of only 1 or 2 stars
out of 5, 90 per cent did not rebook compared with 84 per cent who gave the venue a good rating.
Even among those who had booked the venue at least three times before, a negative review of
the venue reduced the likelihood that they would book again. Of the active customers who reported
a bad experience, 53 per cent did not return during the three years of the study, compared to
48 per cent of those who reported a good experience.
Source: Mitchell, 2014b.
attendance and their postcode, for example, the survey can capture data relating to their attitudes,
opinions and motivations, and subject to the respondent giving permission, the two sources of data can be
linked to give the organization a powerful all-round view of its audiences.
Self-administered surveys
These used to be distributed and collected by placing questionnaires on seats, in programmes, or in other
prominent places in a venue, but are increasingly being done using touchscreen kiosks. As respondents
are self-selecting, bias is likely to creep in through the nature of the sample (see p. 78), but this can
nonetheless be an inexpensive way of collecting data, especially from unticketed events for which
organizations hold few attender contact details, such as galleries or street arts. The number of people who
complete a questionnaire will be significantly improved if it is well explained and laid out, the technology
is easily navigable (if on-screen), the questions are clear and unambiguous, and there is an obvious means
of returning it (if on paper).
Intercept surveys
These are conducted face-to-face with respondents and have several advantages over non-interview
methods:
Interviewers can be involved in the sampling process, selecting appropriate respondents using a
quota system, described above, and eliminating those who are not in the population of interest
They can encourage respondents to answer fully and check for any misunderstandings
Any collateral materials, such as publicity items for respondents to evaluate, can be properly
presented
Interviewers can usually persuade respondents to complete the interview and response rates are
consistently higher than for other types of questionnaire administration
These advantages mean that the quality of the data derived from intercept surveys is generally superior to
other methods, though to some extent this depends upon the skill of the interviewer.
Telephone surveys
This technique has fallen out of favour as online surveys have become more popular and the public has
become more hostile towards unsolicited phone calls. Not all arts organizations systematically store the
phone numbers of their audiences, making systematic sampling difficult, and those who agree to answer
questions on the phone are likely to be those most supportive of a particular cultural organization and may
not be representative of the population of interest. Questions must be relatively simple because it can be
difficult for respondents to understand complex questions over the phone. Interviewers need to be even
more skilled than in face-to-face situations, as non-verbal messages can be neither sent nor interpreted by
interviewer or respondent.
Stage 4: Design the questionnaire

The way in which data are collected will affect the questionnaire length and layout, as well as the style of
questions used. The software being used for analysis will also, to an extent, constrain the style of
questions used, but these constraints can impose useful structure of questions and seldom compromise
the quality of questioning.
Case Study 3.5 Harnessing technology for data collection: Oxford ASPIRE
Two million people were passing through the doors of Oxford’s museums each year, but members
of the Oxford ASPIRE consortium of museums were relying on feedback gained from face-to-face
interviews and paper-based surveys. The insight they could gain into customer opinions was
limited by the time-consuming process of data collection and the relatively small number of
responses that could be collected and processed. So the consortium acquired some free-standing
touchscreen kiosks and mobile survey tablets to capture visitor feedback, enabling visitors to give
their opinions unaided by an interviewer.
A master questionnaire included ten questions relevant to all the museum sites, and each
museum’s touchscreen survey asked additional questions specific to its own services, current
exhibitions and areas of interest. The screens were customized for each museum with logos and
other branding, images and information screens. The Ashmolean also retained a touchscreen tablet
in its shop for running ad hoc surveys around special events, such as late-night openings.
Because the real-time data collected fed into a single database, the different museums could not
only use the system to analyse their own feedback, but could also compare it with the group as a
whole, thanks to a joint museums visitor report which was run quarterly, providing automated
benchmarking.
By using these technologies for data collection, the consortium collected a year’s worth of data
in just three months. As well as collecting significantly more responses than was previously
possible, the system enabled them to be more consistent with their sampling, as they could gather
information throughout their opening hours, rather than being limited to the times when staff were
available to conduct face-to-face surveys.
Among the discoveries they made when they launched the new system was that visitors were
particularly keen on guided tours and access to up-to-date content about collections and museums on
their mobile phones; that the consortium had been more successful at engaging people from harder-
to-reach areas outside the city than it had previously realized; and that two of the museums were more
successful at attracting national and international tourists than others. This knowledge led to them
taking steps to raise the public profile of all the museums, and improve visitor experiences.
Source: Customer Research Technology, 2013.
Criteria for effective questions

As each question is written, it should be evaluated against the following criteria:
Relevance: does it help address the objectives of the survey? If the only questions asked are those
directly relevant to the objectives, the questionnaire length can be kept to a minimum.
Clarity: will respondents understand every word in the question? Complex terminology and ambi-
guity must be avoided at all costs. If respondents fail to understand a question, they will either miss it
out altogether or, worse still, guess at its meaning and give a response to the question they think you
are asking, which may be very different from what you intended.
Brevity: in general, a question should consist of no more than 20 words, otherwise it becomes too
difficult to comprehend quickly.
Impartiality: the wording of the question should not influence the respondents’ answers. This is
more difficult than it sounds. It is quite easy unwittingly to introduce leading questions by failing to
recognize the overtones of certain words.
Precision: each question should deal with a single issue. If two issues are introduced, the respondent
may have opposing views on different parts of the question and may not be able to express these
within the response mechanism provided.
Inoffensiveness: questions relating to sensitive issues, in particular salary, age, ethnicity and social
class, can be construed as offensive unless worded very carefully. They should also be placed at the
end of the questionnaire when the respondent has already committed him- or herself to completing
the questionnaire and is less likely to be put off at this stage.
Question style
There are two basic ways to ask questions in a survey. Open-ended questions allow respondents to
provide answers using their own words, while closed questions (also known as fixed-response or forced-
choice questions) provide a limited number of possible responses which the respondents must choose
between.
Open-ended questions collect information with minimum direction to respondents, and are very
useful if the possible range of responses is very broad. Another advantage of this type of question is that it
uses respondents’ own words, allowing them to vent strong opinions. This can help the researcher to
understand the way that people really think about an issue. Probing questions and clarifying questions
may be used to gain more complete answers to open-ended questions, especially if interviewers are being
used to collect the data. They can build upon subjects offered by the respondents, following up any vague
or general terms given in their answers, though they should offer no prompts or suggestions for cla-
rification as this will introduce bias. The main disadvantage of open-ended questions arises in their
interpretation. All responses have to be read and coded. This means devising a set of categories that
covers the range of all open-ended responses. Inevitably some answers will be ambiguous and difficult to
code precisely, even using software designed for that purpose. Bias can creep in simply from poor coding.
Closed questions offer more limited scope for response, so respondents tend to find them easier and
less time-consuming to complete and most surveys use a high proportion of closed questions to
encourage high response rates. Closed questions also avoid the difficulties of coding, as the code is
written into the question in the form of a response mechanism which prevents respondents producing
ambiguous answers. Despite the limitations on response, closed questions can be used to investigate a
range of issues, including facts, behaviour, attitudes and opinions. They tend to be one of three types:
Multiple choice/response questions
These offer a series of possible responses and can allow people to respond in just one or in a number of
categories. It’s important to include a ‘don’t know’ category for those who are unable to respond. If there
is some doubt as to whether all possible responses have been identified in advance, it’s also important to
provide space for respondents to write in their own answer (the ‘other’ box). For example:
Who booked your ticket? (tick one)
I booked my own á
My partner/husband/wife á
A family member á
A friend á
A colleague á
Don’t know á
Other (please explain) –––––––––– á
Binary questions are multiple response questions where there are only two possible answers, such as
‘true/false’ or ‘yes/no’. In practice, these questions tend to be of limited use because it is often difficult for
respondents to be definite about many answers.
Scaled questions
These provide a scale on which respondents can express the strength of their attitudes or opinions. This
enables different attributes of performances, venues and exhibitions to be compared. Questions can put
forward widely differing opinions of a particular attribute and ask respondents to place their own opinion
somewhere between the polarized views suggested. For example:
Overall, how would you rate this production?
Much better than I expected á

Better than I expected á
As I expected á
Not as good as I expected á
Much worse than I expected á
Likert scale questions ask people to indicate the extent to which they agree or disagree with a par-
ticular statement. They are useful for measuring attitudes. For example:
To what extent do you agree or disagree with the

following statement:
‘Theatres should ban the use of smartphones during
performances’
Agree strongly á
Agree slightly á
Neither agree nor disagree á
Disagree slightly á
Disagree strongly á
Semantic differential questions use words of opposite meaning to suggest descriptions of attributes
of, for example, a venue or performance. Respondents are asked to indicate their opinions on a numbered
scale, for example evaluating the prices of a ticket on a scale of 1 to 10, where 1 is ‘very cheap’ and 10 is
‘very expensive’.
The main problem with scaled questions is that people can be reluctant to express very strong opinions
and therefore tend to avoid the first and last categories in any list. This reduces the sensitivity of the
question as a tool for capturing true attitudes and opinions.
Ordering questions
These enable respondents to express preferences and, if appropriate, to state these preferences in terms of
priorities.
Preference questions will ask respondents for one or more issues or attributes to be identified as being
of more importance than others. For example:
Which three of the following do you feel are the most important
benefits of belonging to the ‘Friends of the Theatre’
Regular information about events á

Ticket discounts á
Discounts on other purchases á
Priority booking á
Exclusive member-only events á
Other (please specify) –––––––––––––––––––––– á
The problem with preference questions is that they tend to magnify perceived differences in pref-
erence. Just because a respondent has excluded an item from the chosen list does not mean that these
items are very unimportant, and the extent to which they are less important is not clear.
Ranking questions take respondents a step further by asking them to prioritize their preferences. In
the example above, respondents could have been asked to rank the options in order, though this proves
difficult for lists longer than just five or six items.
Rankings can be useful, but their limitations must be recognized. They do not indicate strength of
opinion: the item ranked first may be far more important than the second in a respondent’s opinion, but
the second and third may have almost equal importance.
Piloting questionnaires
The pilot stage is when a questionnaire is tested on a few people who are typical of the population
of interest. Questions which seem very straightforward to the researcher may be thoroughly confusing
or even offensive to the respondent, and these need to be identified and changed before the survey
begins.
Stage 5: Collect the data

The data collection method has already been identified by this stage, so the questionnaire is ready to be
administered. Interviewers need to be briefed and trained for face-to-face and telephone data collection,
software configured for online surveys, and technology and administrative systems set up for self-
administered surveys. If the preparation work has been competently undertaken, this can be a very
straightforward part of the marketing research process.
Stage 6: Process the data

Spreadsheets and specialist survey software, such as SurveyMonkey® and SPSS®, can be used to store
and process raw data (i.e. the responses gathered from the sample) ready for analysis.
Data input
If questionnaires are paper-based, the responses from each questionnaire must be typed into the survey
software. This is a time-consuming laborious process, which is bypassed in the case of online surveys and
any data collected on mobile devices.
Data summary
The software will take the raw data and summarize the frequency of responses to each of the questions to
show how many people said what. Visual displays of the data in the form of charts and graphs can depict
this information more graphically. If the sample is representative of the population of interest, then
statistics can indicate how confident one can be that the true figure for the whole population falls within a
narrow band around the observed figure for the sample. Box 3.7 illustrates how the level of confidence is
related to the size of the sample (though, of course, the quality of the questionnaire, the interviewing and
the coding are all factors which will influence the quality of the results, too).
Where appropriate, the software will also calculate the mean (or arithmetic average) response to scaled
questions, and measure the dispersion (or standard deviation) of the result (see Box 3.9). The mean does
not in itself imply that most people thought something. It simply shows the central tendency of the
sample. The standard deviation is calculated to indicate whether the mean was generated as a result of
most people having very similar views, or people having wide-ranging views on a subject so that the
average is simply a halfway position on a very long scale, rather than a representation of popular opinion.
The higher the standard deviation figure, the more wide-ranging were the responses from the sample.
Sometimes the responses fall into two extremes and, if this is the case, it may be possible to use the
findings to identify two discrete market segments within the population.
Data analysis
The process of data analysis uses statistical techniques to identify trends and interrelationships between
the answers to different questions. Cross-tabulation is a useful approach. This places data into two-way
tables, enabling responses to any question to be compared with the responses to any other question. For
example, using cross tabulation it would be possible to conclude whether sitting in the stalls or in the gods
affects the extent to which the audience enjoys a concert or play. Or the expectations of tourists from a
visit to a gallery could be compared with the expectations of local people, school groups and art students.
Any observed differences that emerge in the tables can then be investigated more closely using a wide
range of statistical tests. While these tests can be performed quickly and easily using survey software, the
interpretation of the numeric output requires a high level of statistical expertise.
Box 3.9 Mean and standard deviation

Suppose 20 people were asked the extent to which they agreed or disagreed with the following
statements:
Statement 1 Statement 2
‘Children under the age of 14 should not be allowed into art galleries ‘Museums should never charge
unless accompanied by an adult.’ entrance fees.’
1 = strongly agree 4 = disagree slightly
2 = agree slightly 5 = strongly disagree
3 = neither agree nor disagree
The different responses of the 20 people to the two statements might be as follows:
Statement 1
1, 5, 4, 5, 2, 2, 5, 5, 3, 3, 1, 1, 4, 4, 1, 2, 4, 5, 2, 1
Statement 2
2, 4, 4, 3, 3, 4, 1, 2, 4, 4, 3, 3, 1, 5, 4, 2, 2, 3, 4, 2
8
Statement 1
7 Statement 2
6
Number of responses
0
1 2 3 4 5
Level of agreement
Figure 3.7 Statements.
Statistically, these results indicate an identical average response to each statement but different
dispersions:
Statement 1 Statement 2
mean = 60/20 = 3 mean = 60/20 = 3

standard deviation = 1.59 standard deviation = 1.13
The relatively high standard deviation in the responses to Statement 1 implies that people have
stronger feelings about the admission of lone children to art galleries, either strongly in favour or
strongly against. The mean, in this case, does not imply that the majority of people have neutral
feelings on the subject, simply that firmly held attitudes are evenly balanced. There is less
divergence in Statement 2, implying that attitudes are more consistent throughout the population.
(Note: this example is given only to illustrate the mean and standard deviation; in reality the
sample size would be too small to make judgements about the nature of the population.)
Stage 7: Interpreting and reporting findings

The final step of a survey is to translate the analysis into usable information which can support marketing
decisions.
Interpretation
Figures do not speak for themselves. It is their interpretation that renders them useful (or useless).
In interpreting data, a number of key points must be constantly borne in mind:
The interpretation should relate to the problem identified at the very beginning of the survey process,
and the only findings which should be highlighted should be those which relate to the explicit
objectives of the research.
Statistics are not definitive and few figures are absolute. Most statistics indicate the likelihood that a
population will behave in a certain way, or hold a certain opinion. This should be recognized in the
way that the figures are used to draw conclusions.
Findings should not be exaggerated or dramatized to draw attention to them. Fair weight should be
given to all the relevant evidence.
Small sample sizes should be recognized and generalizations should not be made based on samples
which are too small to give statistically significant results.
Averages should be recognized as merely tendencies and not be thought of as representing a ‘Mr and
Mrs Average’.
Recognition should be given to infrequent but relevant answers to open questions, as they may
indicate an emerging trend.
There should be clear distinction between fact (what people are, what they do, what they own etc.)
and opinion (what people think about something). Opinions and attitudes do not always translate
into behaviour.
There should be clear distinction between cause and effect. Associated factors are not necessarily
causal. Surveys are useful for identifying effects, but conclusions about causes must be drawn very
tentatively and sometimes need further investigation.
Reporting and presentation

If a report or presentation is to be used to communicate survey findings to decision makers, their needs
must be its focus. There are four main points to bear in mind.
Present only information that is relevant to the reader
Box 3.10 Devil in the detail

The English government’s Department for Culture, Media & Sport (DCMS) was quick to
celebrate the success of its policy of offering free admissions to the country’s major museums
and galleries. Figures revealed that 47 million visits were made to the 16 museums and
galleries directly funded by British taxpayers in 2014 – a record high since the first data was
published.
But the DCMS was slower to point out how the profile of those visitors had changed. Rather than
encourage domestic attenders to use their country’s museums and galleries, closer analysis of the
data revealed that this funding had been much more successful at attracting visits by foreign
tourists. While international visitor numbers grew by almost 40 per cent since 2008–2009 when
comparable records began, visitor numbers from within the UK increased by just 3 per cent during
this same period, a time when population growth itself was over 4 per cent.
Source: ArtsProfessional, 2015.
Too much information is confusing, but there must be enough detail to guide any decision that is to be
made. For example, it is important to give information on sample sizes and details of the research
methodology, to enable the reader to estimate the accuracy of the findings. A copy of the questionnaire
should also be included to allow the reader to refer back to questions posed. Some details can be left out.
There is no need, for example, to list all the responses to open questions. Also, if a large number of small
frequency responses are given to a closed question, some of the categories may be merged and reported
under the heading ‘other responses’.
Communicate the findings simply
Report findings visually, highlighting only the most important parts in text. Tables and graphs are par-
ticularly useful for highlighting important features or trends. Technical jargon should be kept to a
minimum and simple statements used to convey the main points.
Do not imply certainty
No survey is absolutely definitive and no figure in a survey is an entirely accurate representation of

reality. Therefore findings should often be reported as whole percentages rather than to decimal places,
to avoid implying a misleading level of precision. If samples or sub-samples are small, readers should
be warned that the findings should be treated with caution. Alternatively, the confidence interval (see
Box 3.7) for the reported percentages should be given.
Provide an executive summary
Not every reader will want to read the whole report, so a brief summary of the findings, the main
conclusions and recommendations should appear at the beginning. Complicated tables, discussion of
methodology and analysis should be kept for other sections of the report.
Conclusion
Much of the theory of marketing research applies to arts organizations in exactly the same way as it does
to commercial organizations. A desire for more information needs to be formalized by setting research
objectives, and the gathering and interpretation of facts and figures should be a structured and methodical
process. The financial constraints facing many arts organizations don’t prevent them conducting mar-
keting research, but simply means that they must be especially focused in the objectives they set and
resourceful in the techniques they use. The more focused the research is, the more likely it will be of
practical value to the organization.
Marketing research is like a taxi-cab – it will go anywhere you want (providing you have the fare) but
the cab driver must know how to drive, must have a working knowledge of the Highway Code, and the
passenger and the driver must agree at the outset as to the final destination (Webb, 1993).
End of chapter discussion case study

3.6 Investigating the user experience: the Broad Art Museum
The Eli and Edythe Broad Art Museum is a contemporary art museum situated on the Michigan
State University campus in the USA. The Broad is keen to be closely connected to its community
and to encourage its visitors to engage fully with the museum – as ‘cultural participants’, rather
than passive consumers.
To do this, they first needed to understand how people from different communities interacted
with their museum, and what influenced those interactions. But gathering this data would not be
easy, for two reasons. Firstly, free admission meant that many people passed through without
ever interacting with staff, making it difficult even to find out about the demographic profile of
visitors. The museum wanted to know more about these visitors, but without disrupting their
experiences at the museum. Secondly, the Broad has entrances and exits on the east and west
sides of the museum, but no single front door, so visitors can easily slip in and out of the building
anonymously. Gaining a better understanding of the visitor experience was difficult because of
the museum’s layout.
Planning
To understand how best to tackle these challenges, a group of researchers and museum
practitioners set about designing a research project that would enable the museum to decide
how best to gather information about the visitor experience without frustrating people. They
identified three key research issues:
How to collect relevant information on the visitor experience

In what format to collect the information so as not to inconvenience visitors
At what point during a visit to collect the information
Three types of research were planned to help make these decisions: contextual observation,
intercept surveys and interviews.
Contextual observation
To understand how visitors entered and left the museum, and how they interacted and participated
while they were there, visitors were observed for two hours each day for a week. Observers did a
headcount every 15 minutes to get a sense of how many people were using the different areas of the
museum. This research had some limitations, as the architecture of the Broad made it difficult in
places for observers to see how people interacted with exhibits.
Nonetheless, this research revealed that visitors generally started and ended their visits in the
social spaces around the museum cafe and gift shop. It also suggested that while students, lecturers
and others associated with the university used the museum in fairly similar ways, this was not true
of families, who appeared to use the Broad less often as a social space, and more often to view
exhibits.
Intercept survey
These findings led to the research questions being revised to focus specifically on how families
interacted with exhibits and resources, and to understand their level of satisfaction with their
overall experience.
Interviewers used touchscreen tablets to administer a questionnaire that could be
completed in less than 30 seconds as families left the museum. It was designed to find out
how families would describe their experience after visiting the Broad, and asked them to use
a single word or short phrase to describe their experience. Thirty-nine families were
interviewed over two days.
The Broad’s website is slightly difficult to Travelling to and from the Broad does not Guests are intrigued by the building itself Opportunities to participate are excellent,
navigate on mobile and on desktop. appear to be an issue for visitors. and, overall, enjoy their visit. but there is room for growth.
Stages Planning/Deciding Researching Travelling Visiting Leaving Post-visit
Looking for a family Searching the internet; Drive to the Broad and Viewing and interacting Finding car to drive to Looking for information
Broad activity on a weekend. talking with others. nd
fi parking. with site and materials. the next destination. about activities for family.
Gathering
Information Website Walking Reflecting
Social Find parking on campus
Media
Planning a trip to
Driving participate in a
Doing
Walk through the museum community event
Deciding
Participated in exhibits, but most
!
Viewing Only aware of
time was spent in the education Outdoor minor exhibits on
area for families to read and Exhibits the west side
Colleagues Friends explore artwork.
• What family activities do they have? • Did we come through the correct entrance? • Did we miss any of • Plan to go back on a
Thinking • Do they have snacks for the kids? • What exhibits are appropriate for children? the exhibits? rainy or cold day?
• Want to find out more

• Interested in the exhibits • Peaceful and quiet inside the museum • Excited to plan a visit
Feeling • Curious about the other community events • Desire more participatory exhibits for kids
about family activities
on a family day
offered by the Broad
• Spent majority of time in children’s area • Unaware of outdoor • Researching more

• Website is somewhat confusing to navigate
Experience • The Social Media pages are more clear
• Children’s area may be too small for multiple exhibits on the east side information about
families to utilize at once and did not view them family days
Website could perhaps be updated more often. Create a larger space for children. The space could Encourage museum personnel to engage guests
Opportunities “The last update for the first family Saturday of the encourage kids and families to create their own with outdoor exhibits. Provide family activity
month was way out of date.” artwork to decorate this space. information to families during their visit.
Figure 3.8 Experience map of a family’s visit to the Broad (from Campbell et al., 2016).
Source: Campbell et al., 2016. © UXPA 2016. Reprinted from User Experience, Vol. 16, No. 4, September
2016. http://uxpamagazine.org/solving-a-museums-businesschallenges
Visitors came from postcodes across the whole of Mid Michigan, 74 per cent of them first-time
visitors. Over half had heard about the Broad through word of mouth and 71 per cent said they were
very likely to visit the Broad again. It was the content of the exhibits, and not the end-to-end
experience of visiting the Broad, that tended to affect how they rated their overall experience.
Depth interview
To learn about the context for some of the statements made on the survey, and assemble an
‘experience map’ of an end-to-end visit to the Broad, it was decided to conduct between four and
six depth interviews. Survey respondents had been asked whether they would be willing take part
in an interview, and those who had ticked the box were emailed an invitation to participate. The
researchers also searched among their own networks for families that were first-time Broad
visitors, but despite this, it was only possible to recruit one family comprising two adults and two
children.
The goal for the interview was to learn about what the family was doing, thinking and feeling as
they visited the museum. Researchers drafted the map, and the family was invited to edit language
or visuals, to ensure that it was a fair representation of their experience (see Figure 3.8).
As well as illustrating key findings that emerged from the observations and the survey, such as
the visit ending at the cafe and gift shop common area, it identified some opportunities for
improving the end-to-end experience for visitors. Website navigation was found to need attention,
and there were some useful suggestions about how to improve the experience for their children.
The map also raised questions about how families interact with social media, the Broad website,
and the outdoor exhibits.
What next?
At the end of the project, the research team concluded that the project had given them a deeper
insight into the way visitors interact and experience the exhibits at the Broad. They planned to
extend their research project to design and test creative interactive ways of collecting demographic
data on visitors in the Broad’s social spaces, such as the cafe, gift shop and education space, in a
way that would add value to visitors’ experiences at the museum in ethical and minimally
disruptive ways.
Source: Campbell et al., 2016.
Questions
1. To what extent did the Broad achieve its research objectives?
2. Discuss the limitations of the sample. What strategies could you use to maximize visitor
response to a survey?
3. Draft a proposal for the Broad, suggesting how it could conduct visitor research in a way that
would enable it to collect more demographic data in ethical and minimally disruptive ways,
adding value to visitors’ experiences at the museum.
Note
1 The term ‘marketing’ research is often shortened to ‘market’ research, though strictly speaking this is incorrect.
Marketing research is a wide-ranging discipline which seeks to gather information about any aspect of an
organization and its environment, while market research is concerned only with trends in customer or audience
behaviour, and is usually known as ‘audience research’ in an arts context.
Bibliography
Albert, S. (2016) Measuring aesthetic value. ArtsProfessional. Available from www.artsprofessional.co.uk/
magazine/298/feature/measuring-aesthetic-value. [Accessed 1 January 2017]
Anderson, K. (2013) A Culture of Research – Bringing Research Into Your Day-to-Day Practice. Available from
http://culturehive.co.uk/wp-content/uploads/2014/01/KatieAnderson_WarwickArtsCentre_aCultureOfResearch.
pdf. [Accessed 1 January 2017]
ArtsProfessional. (2015) Culture budget pays for tourism agenda. Available from www.artsprofessional.co.uk/
news/culture-budget-pays-tourism-agenda. [Accessed 1 January 2017]
ArtsProfessional. (2016) Metrics to measure arts quality? The sector speaks out. Available from
www.artsprofessional.co.uk/magazine/article/metrics-measure-arts-quality-sector-speaks-out. [Accessed 1 January
2017]
Axelsen, M. (2006) Using special events to motivate visitors to attend art galleries. Museum Management and
Curatorship. Vol. 21, No. 3. Available from www.tandfonline.com/doi/citedby/10.1080/09647770600302103?
scroll=top&needAccess=true. [Accessed 4 October 2016]
CACI Information Services. (2016a) Acorn Technical Guide. Available from http://acorn.caci.co.uk/downloads/
Acorn-Technical-document.pdf. [Accessed 7 October 2016]
CACI Information Services. (2016b) Acorn User Guide. Available from http://acorn.caci.co.uk/downloads/
Acorn-User-guide.pdf. [Accessed 7 October 2016]
Campbell, E., Lauren, B., Stoepel, W. & Kribs, S. (2016). Solving a museum’s business challenges: A case study.
User Experience Magazine. Vol. 16, No. 4. Available from http://uxpamagazine.org/solving-a-museums-
business-challenges/. [Accessed 31 October 2016]
Culture24. (2014) Tracking Your Audience’s Online Motivations. Available from www.keepandshare.com/doc/
7398823/tracking-your-audiences-online-motivations-pdf-769k?da=y. [Accessed 31 October 2016]
Customer Research Technology. (2013) Aspiring to Deliver Visitor Experience Dxcellence. Available from
www.crtviewpoint.com/sites/default/files/oxford_aspire_case_study_august_2013.pdf. [Accessed 13 October
2016]
Department for Culture, Media & Sport. (2016) Taking Part 2015/16 Quarter 4 Statistical Release. Available from
www.gov.uk/government/uploads/system/uploads/attachment_data/file/539312/Taking_Part_2015-16_Quarter_
4_Report_-_FINAL.pdf. [Accessed 2 January 2017]
Evans, R. & Tomlinson, R. (2013) Experiments: Exploring and Testing Arts Marketing Assumptions Every Day.
Conference Report, Arts Marketing Association. Available from http://culturehive.co.uk/wp-content/uploads/
2014/01/Experiments_ExploringAndTestingArtsMarketingAssumptionsEveryDay1.pdf. [Accessed 2 January
2017]
Experian Limited. (2014) Mosaic: The Consumer Classification Solution for Consistent Cross-channel Marketing.
Available from www.experian.co.uk/assets/marketing-services/brochures/mosaic_uk_brochure.pdf. [Accessed
2 January 2017]
Falk, J. (2009) Identity and the Museum Visitor Experience. Walnut Creek, CA: Left Coast Press.
Fantoni, S. F., Stein, R. & Bowman, G. (2012) Exploring the relationship between visitor motivation and engage-
ment in online museum audiences. Museums and the Web 2012 11–12 April 2012, San Diego, California USA.
Silver Spring, Maryland: Museums and the Web. Available from www.museumsandtheweb.com/mw2012/
papers/exploring_the_relationship_between_visitor_mot. [Accessed 3 October 2016]
Fontein, D. (2016) The Top 26 Social Media KPIs Marketers Can’t Ignore. Available from https://blog.hootsuite.com/
social-media-kpis-key-performance-indicators/. [Accessed 11 October 2016]
Gribling, L. (2014) How to Manage Data and Research to Support Fundraising. Available from http://culturehive.co.uk/
wp-content/uploads/2014/09/Rotterdam-Philharmonic-fundraising-case-study.pdf. [Accessed 3 October 2016]
Higgins, C. (2016) Orwellian nightmare or fairer system for all? What Quality Metrics will mean for arts funding.
The Guardian. Available from www.theguardian.com/culture/2016/oct/04/quality-metrics-arts-council-england-
funding [Accessed 2 January 2017]
Humphreys, B. (n.d.) How Acorn Segmentation of Audiences can Help Arts Marketers Target their Communi-
cation. Available from http://culturehive.co.uk/resources/how-acorn-segmentation-of-audiences-can-help-arts-
marketers-target-their-communication. [Accessed 7 October 2016]
Isaacson, B. (2012) Why You Should (Almost) Never Use the van Westendorp Pricing Model. Available from
http://mmrstrategy.com/why-you-should-almost-never-use-the-van-westendorp-pricing-model. [Accessed 2
January 2017]
Lewis, A. (2014) What Can we Learn from Watching Groups of Visitors Using Digital Museum Exhibits? Available
from www.vam.ac.uk/blog/digital-media/digital-exhibits-observational-research. [Accessed 9 October 2016]
Mitchell, H. (2014a) Enjoying the Christmas rush. ArtsProfessional. Available from www.artsprofessional.co.uk/
magazine/278/article/enjoying-christmas-rush. [Accessed 7 October 2016]
Mitchell, H. (2014b) Seizing the moment. ArtsProfessional. Available from www.artsprofessional.co.uk/magazine/
277/article/seizing-moment. [Accessed 12 October 2016]
Moriarty, G. (1997) Taliruni’s Travellers: An Arts Worker’s View of Evaluation. The Social Impact of the Arts
Working Paper 7. Stroud: Comedia.
Tasich, T. & Villaespesa, E. (2012) Come along with us. Journal of Arts Marketing, Vol. 48, pp. 17–20. Available
from http://culturehive.co.uk/wp-content/uploads/2013/02/Article-Come-along-with-us-JAM48-Tijana-Tasich-
Elena-Villaespesa-2012.pdf. [Accessed 2 January 2017]
The Audience Agency. (n.d.) Audience Spectrum Pen Portraits. Available from www.theaudienceagency.org/
audience-spectrum/profiles. [Accessed 26 March 2017]
Tomlinson, R., Roberts, T. & Allpress, V. (2006) Full House: Turning Data into Audiences. New Zealand: Creative
New Zealand.
TNS BMRB. (2016) 2015/16 Taking Part Survey Technical Report. Available from www.gov.uk/government/
uploads/system/uploads/attachment_data/file/543460/Taking_Part_Technical_Report_2015-16.pdf. [Accessed
2 January 2017]
Ward, P. (2016) Michelangelo: ‘Not Ticked All the Metrics’. Available from http://fundermental.blogspot.co.uk/
2016/09/michelangelo-not-ticked-all-metrics.html. [Accessed 2 January 2017]
Webb, J. R. (1993) Understanding and Designing Marketing Research. London: Academic Press.
4 Product
Product is arguably the most important element of the four ‘Ps’ of product, price, promotion and place
which constitute the marketing mix. A product can be either a service (like a ride in a taxi) or a physical
good (like the taxi itself). It is essentially a bundle of customer benefits offered for sale. It can consist of
tangible or intangible elements, or – more commonly – a combination of both. Some writers prefer to use
the term ‘product’ to denote physical goods as opposed to services, but in this chapter we will use it as an
inclusive term covering both.
Product provides the basic building block of any marketing strategy, and has a central role in arts
marketing because of the innovative and creative nature of artistic inspiration. But what concerns the arts
marketer is not so much what is produced by artists as what is available to audiences as experience. This
chapter will address the nature of the experience of art from a customer perspective, exploring how
marketing can optimize and enhance outcomes for customers and arts organizations. We will discuss:
The nature of goods and services

Service-dominant logic
Levels of product
Product decisions
Service quality
Goods and services

Marketing theory and practice has been built largely on the experience of companies selling packaged
goods in the 1950s and beyond. But arts organizations (along with other service-providing organizations
such as banks, doctors and hairdressers) create benefits for customers by selling services rather than
physical goods. Services have characteristics which physical goods lack, and which affect how they are
marketed. Four widely recognized ones are:
Intangibility
Inseparability of production and consumption
Heterogeneity
Perishability
Intangibility
Unlike physical goods, which can be handled and stored, services are intangible. They are experiences
rather than objects. Potential consumers cannot inspect a night out at the theatre before purchase in the
same way as they might, for example, test-drive a car.

Cap3 - MKT 2022-11-18 13 - 56 - 53

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cap3 - MKT 2022-11-18 13 - 56 - 53

Uploaded by

Copyright:

Available Formats

3 Marketing research

 The benefits and scope of marketing research

The benefits of marketing research

Objections to marketing research

‘ : : : it’s too expensive’

‘ : : : we don’t want to be popular’

‘ : : : we know what our audiences want’

Box 3.1 Quality Metrics

The scope of marketing research

Customer satisfaction research

Case Study 3.1 Web motivations: Indianapolis Museum of Art

Planners: Planning a visit to the museum (50 per cent)

Box 3.2 Measuring price sensitivity

Case Study 3.2 Taking Part in culture

Conducting marketing research

What do we want to know?

How can the information be obtained?

How will we use the findings?

Why set objectives?

Box 3.3 Eavesdropping at Tate Modern

 Motivations and inhibitions for participating in the arts

Sources of secondary data

Box 3.4 Key performance indicators

 Why should this performance indicator be measured?

Category Key performance indicator

Reach and brand Visits and searches

Case Study 3.3 Christmas bookers

Up to and October November December December January

Figure 3.2 Pre- and post-opening statistics.

Existing Booker New Booker

Figure 3.3 2012–2013 graph.

Geodemographic profiling systems

Trips and Treats

Home and Heritage

Figure 3.4 Audience Spectrum (from The Audience Agency, n.d.).

Box 3.5 Finding flourishing families

Prevalence among venue attenders

Prevalence in High Average Low

Figure 3.5 Prevalence in local population (from Humphreys, n.d.).

Collecting primary data

Case Study 3.4 The value of digital exhibits

Box 3.6 Testing times

Figure 3.6 The stages of ad hoc survey research.

Stage 2: Identify an appropriate sample

 When should I sample?

 Who should be in my sample?

Box 3.7 Sample sizes

Table 3.1 Sample sizes

Sample size Response frequency (%)

100 5.9 7.8 9.0 9.6 9.8

Stage 3: Decide how the survey will be implemented

Box 3.8 Did you enjoy that?

Stage 4: Design the questionnaire

Criteria for effective questions

 Multiple choice/response questions

Who booked your ticket? (tick one)

Overall, how would you rate this production?

Much better than I expected á

To what extent do you agree or disagree with the

Regular information about events á

Stage 5: Collect the data

Stage 6: Process the data

Box 3.9 Mean and standard deviation

The benefits and scope of marketing research

Motivations and inhibitions for participating in the arts

Why should this performance indicator be measured?

When should I sample?

Who should be in my sample?

Multiple choice/response questions

Present only information that is relevant to the reader

Communicate the findings simply

Do not imply certainty

Provide an executive summary

How to collect relevant information on the visitor experience

The nature of goods and services